In an astonishing turn of events, Apple has secretly told the press that iPhone X facial recognition will not work well. So you can pre-order an iPhone X on Friday if you want. But you may want to hold off a bit.
I assume you’re up on the self-imposed issues that Apple faces at the moment. But the short version is this: The Cupertino consumer electronics giant misfired on a new iPhone strategy this fall by launching three new models for the first time. Two of them, the iPhone 8 and 8 Plus, are old-fashioned looking and have faced battery reliability issues. And one of them, the iPhone X, is more forward-leaning in aping the near bezel-less screen design that Samsung innovated, but with a compromising “notch” at the top of the screen that even Apple’s biggest fans have described as a mistake.
Sign up for our new free newsletter to get three time-saving tips each Friday — and get free copies of Paul Thurrott's Windows 11 and Windows 10 Field Guides (normally $9.99) as a special welcome gift!
"*" indicates required fields
I think most would forgive Apple the notch were it not for the iPhone X’s suspicious reliance on facial recognition for log-in. Unlike the previous several generations of iPhones, the iPhone X does not feature a Touch ID sensor for finger-based sign-ins. And with Apple unable to get in-screen fingerprint detection working in time for the launch, the firm went with plan B: Facial recognition.
Well, now it’s going with plan C. Which is to seed the press with the bad news that this technology does not work very well. In doing so, it can temper expectations for the product and assure that only its most-forgiving fans will buy an iPhone X, preventing the public embarrassment of rampant complaints.
“Apple quietly told suppliers they could reduce the accuracy of the face-recognition technology to make it easier to manufacture, according to people familiar with the situation,” Bloomberg reported. “A less accurate Face ID will still be far better than the existing Touch ID [but] the company’s decision to downgrade the technology for this model shows how hard it’s becoming to create cutting-edge features that consumers are hungry to try.”
I’m sorry, what? “A less accurate Face ID will still be far better than the existing Touch ID”? That cannot be true. Touch ID is fantastic. Even a full-working Face ID would likely not be as good.
Anyway, the Bloomberg report goes on to chronicle the Apple-centric history of what happened with iPhone X production: Apple was struggling to get sufficient components for the phone and needed fewer people to put it together. And the main culprit, its source say, was the 3D sensor that recognizes faces and unlocks the handset.
Apple’s “tight schedule” further complicated matters, Bloomberg says. It underestimated the complexity of making and assembling exceedingly fragile components, the sources told it. And this also explains why the iPhone X is shipping six weeks after the iPhone 8/Plus.
As you may know, the facial scanner in the iPhone X is based on the technology that Microsoft first used, disastrously, in its Xbox Kinect sensor. This probably explains why it works so poorly: If Microsoft could never perfect this in a relatively huge device, how could Apple’s component makers ever fit the technology into “a space a few centimeters across and millimeters deep”?
This has the makings of a disaster, and in the sense that it’s always tough to bet on first-gen technology, you might want to just take a year off on this one. I’m sure the iPhone XS (“excess”) will get it right in late 2018.
And Bloomberg has come to the same conclusion I have: Apple will likely sell fewer handsets in a launch quarter than it did the year before for the first time ever.
“Signs of weakness in iPhone 8 sales means Apple could sell fewer handsets than last year, despite all the fanfare surrounding the iPhone X,” it notes. Yep.
skane2600
<blockquote><a href="#210255"><em>In reply to Daekar:</em></a></blockquote><p>You don't start relaxing specifications just weeks before you start selling a new product unless you have a really good reason. Actually there's probably never a good reason. You wait until your product has been used by real-world customers for awhile and then you see if you can adjust the specs. </p>
skane2600
<blockquote><a href="#210282"><em>In reply to unkinected:</em></a></blockquote><p>We don't know the exact timeline, but whether it's a last-minute change before launch or a last-minute change during initial manufacturing the fundamental risk is the same. Clearly if the iPhone X was "always going to launch this way" there would have been no need to communicate a change in specs.</p>
skane2600
<blockquote><a href="#210284"><em>In reply to Bob_Shutts:</em></a></blockquote><p>Because if a company denies a story it cant be true? In any case, the one in a million figure is obviously made up.</p>
skane2600
<blockquote><a href="#210466"><em>In reply to PincasX:</em></a></blockquote><p>If you think carefully about how you could come up with such a number, you'd get it. It would involve knowing how much people's facial characteristics vary across the entire world, the accuracy of the sensor and the efficacy of the software algorithm that determines whether there's a match. Of the three, only the accuracy of the sensor can be measured as a practical matter. There may be other factors that would affect the accuracy of the prediction.</p>
skane2600
<blockquote><a href="#210482"><em>In reply to PincasX:</em></a></blockquote><p>Once it's been demonstrated that the number can't be arrived at scientifically what explanation is left?</p><p><br></p>
skane2600
<blockquote><a href="#210489"><em>In reply to PincasX:</em></a></blockquote><p>I believe I just did. Or do you think Apple has done a 3D scan of everyone in the world?</p><p><br></p><p>Here's Apple's Face ID security document. Notice that they don't say anything about how they came up with the 1 in 1,000,000 probability, they just state it as a fact:</p><p>images.apple.com/business/docs/FaceID_Security_Guide.pdf</p><p><br></p>
skane2600
<blockquote><a href="#210607"><em>In reply to PincasX:</em></a></blockquote><p>You can theorize all you want about bone structure negating the need to study all faces (or even a statistically significant percentage) but it's still just a theory, not a proven fact. Yes, scientists can look at a skull and tell you something about that individual's appearance but not necessarily about anyone else's appearance. If you can point to some scientific paper that claims that facial features of everyone in the world (or again a statistically valid subset) is known based on some common features, I'm all ears.</p><p><br></p><p>You should also understand that Apple's system uses neural nets whose method of determining an outcome in unknowable. While a conventional algorithm can be analyzed to determine is efficacy, the efficacy of a NN can only be approximately determined by testing it against a large amount of data. http://www.technologyreview.com/s/604087/the-dark-secret-at-the-heart-of-ai/</p><p><br></p><p>In conclusion, the lack of knowledge about facial features across the entire world combined with lack of understanding of how the "algorithm" works means that no credible estimation of correctness can be determined. </p>
skane2600
<blockquote><a href="#210740"><em>In reply to PincasX:</em></a></blockquote><p>There's not much conjecture in my claim. I have a 25-year-old Neural Network textbook on my shelf that says the same thing as the article I referenced about the inability to know how a NN comes up with an answer. You probably disagree, but IMO anyone claiming that Apple has enough data on people's faces to make a determination of the accuracy of their face recognition system has more of a burden of proof than anyone who disputes it (even if the inscrutability of neural nets were not a factor).</p><p><br></p><p>If MS, Google, or Apple announced a probability of a bug in their software or hardware would you believe it? Why is this particular claim of Apple's any different?</p><p><br></p><p>Finally, keep in mind that I'm not making any claims about the efficacy of their system, for all I know it could be more accurate than 1 in 1,000,000, I'm just questioning their ability to estimate that efficacy with reasonable accuracy.</p>
skane2600
<blockquote><a href="#210823"><em>In reply to PincasX:</em></a></blockquote><p>I haven't changed my position at all. I was justifying my conclusion that it was "obviously made up" on the basis that it was impossible for Apple or anyone else to prove their 1 in 1,000,000 claim. You were certainly presenting arguments to support Apple's claim. although it's true you didn't declare their numbers as valid. For someone who had no opinion on the merits of the claim, you certainly had a lot of things to say, trying to refute some of my arguments while ignoring others. </p>
skane2600
<blockquote><a href="#210884"><em>In reply to PincasX:</em></a></blockquote><p>Having failed to prove your point you now resort to insults. If you think I'm a cheerleader for Paul you must be new around here. I've posted comments that disagree with him on many occasions. </p>
skane2600
<blockquote><a href="#210885"><em>In reply to PincasX:</em></a></blockquote><p>I've made no claim that Paul's article was true, so I don't know why you're bringing that up. </p><p>Some conclusions don't require data to justify them, just an understanding of certain logical and domain-specific knowledge . For example, I don't need to see the plans for any particular perpetual-motion machine to know none of them will work (I'm not comparing such machines to Apple's facial recognition system, I'm just illustrating a general principle.)</p><p><br></p><p>You simply cannot prove the probability of an error in any complex system with significant precision. It's a limitation that doesn't uniquely apply to Apple. If I saw Google, or Microsoft or any other tech company make such a claim, I'd call them out too.</p>
skane2600
<blockquote><a href="#210922"><em>In reply to PincasX:</em></a></blockquote><p>I assumed you had a rational (if flawed) reason for referring to "Pom poms". My mistake, it was merely a random insult. </p>
skane2600
<blockquote><a href="#211035"><em>In reply to PincasX:</em></a></blockquote><p>I wrote many posts justifying my claim that Apple made their numbers up (I didn't say "lie", BTW), and you think it's equivalent to your unexplained Pom Pom insult. LOL.</p>
Stooks
<blockquote><a href="#210306"><em>In reply to GT_Tecolotecreek:</em></a></blockquote><p>You are not any better than Paul. </p><p><br></p><p>You have no proof that iPhone X will sell well. You also have zero proof that poor iPhone 8 sales are because of people are waiting for the X. </p><p><br></p><p>The iPhone 7 is out selling the 8 according to something I read over on Macrumors based on some sales data. That means people are still buying the 7 more than the 8 but people are buying the 8 as well. That is a lot of people buying iPhones right now. Who is waiting for the X?</p><p><br></p><p>So is your guessing better than Paul's???</p>
Stooks
<blockquote><a href="#210509"><em>In reply to monkeyboy:</em></a></blockquote><p>I agree and the shrinking market share of both iPhones and Mac's should be a concern for Apple. Last I read Android now has 85%, iPhone 14% and 1% for everything else. Also Mac market share was at 9.5% in April of 2016 according to Netmarketshare, and was down to 5.9% in September of this year.</p><p><br></p><p>All that said I would not bet against Apple yet. They are not after market share so much. They focus on profit margins big time. That said with shrinking market share they are forced to raise prices. How long can that strategy last? </p><p><br></p><p>I have lot's of iPhone using friends but none of them are going to pay $1000 for a phone. They will either go after the "$549" target which is the 32gig iPhone 7 right now or move to Android at some point. What is surprising from the people I know is that they pretty much don't care anymore. Smartphones are becoming so common place and the abilities so equal across the board that the wow factor is gone. It is a device that does a specific set of things for them. The hardware in terms of capability is pretty much equal when it comes to those common tasks and same goes for Android or iOS at this point.</p>
junjunralriosa
<p>Tradingis the best thing that ever happened to me. I'm a student and I'm proud to say that I'm profiting really well. If you want to learn how totrade and profit fromtrading. Search SuperiorTrading System.</p>